As the capabilities of large language models (LLMs) emerge, they not only assist in accomplishing traditional tasks within more efficient paradigms but also stimulate the evolution of social bots. Researchers have begun exploring implementation of LLMs as the driving-core of social bots, enabling more efficient and user-friendly completion of tasks such as social behavior decision-making and social content generation. However, there is currently a lack of systematic research on behavioral characteristics of LLMs-driven social bots and their negative impact on social networks. We have curated data fromChirper.ai, a Twitter-like social network populated by LLMs-driven social bots and embarked on an exploratory study. Our findings indicate that: 1) LLMs-driven social bots possess enhanced individual-level camouflage while exhibiting certain collective characteristics; 2) these bots have the ability to exert influence on online communities through toxic behaviors; and 3) existing detection methods are applicable to LLMs-driven social bots but may have certain limitations in effectiveness. Moreover, we organized the data collected in our study into Masquerade-23 dataset, which we have publicly released, thus addressing the data void in subfield of LLMs-driven social bots behavior datasets. Our research outcomes provide primary insights for the research and governance of LLMs-driven social bots within the research community.

The inherent scarcity of high-quality electroencephalography (EEG) datasets critically constrains the development of robust brain-computer interface (BCI). Data augmentation has thus emerged as a crucial strategy for artificially enlarging the dataset. However, existing augmentation frameworks often struggle to generate highfidelity signals. In this paper, we propose a novel EEG data augmentation framework based on the Graph-Embedded Conditional Diffusion model for generating artificial EEG (GCDE) to augment dataset and improve the performance of EEG decoder. Unlike Variational Autoencoders (VAE) and Generative Adversarial Networks (GAN), GCDE employs an iterative denoising process to generate realistic signals. We integrate Graph Embedding within a U-Net architecture to learn the spatial topological relationships among multiple EEG electrodes, thereby capturing the complex temporal and neuroscientific significance inherent EEG signals. Additionally, we incorporate label embedding to enable conditional, classspecific generation. We evaluate GCDE on the BCI Competition IV 2a dataset and investigate the optimal ratio that maximizes performance improvements. GCDE achieves significant improvements of 6.95 % in EEGNet and 3.36% in the ST-GF model when the generated data constitutes 75 % of the real data. To validate its generalizability, we further conduct experiments on the SEED (emotion recognition) and Fatigue (fatigue detection) datasets, where accuracy with EEGNet increase to$98.87 \%(+2.64 \%)$and$93.81 \%(+3.60 \%)$, respectively. These comprehensive results demonstrate that GCDE is a powerful and generalizable framework for generating high-quality EEG signals, advancing data augmentation in BCI and other EEGbased domains. Our code will be released upon acceptance.

Abstract The privacy of irrelevant individuals appearing in social media videos is often compromised, which presents a challenging issue for privacy protection. Current approaches to privacy protection primarily rely on manually identifying irrelevant individuals, which is both time-consuming and labor-intensive. Therefore, this paper proposes an automatic algorithm for identifying irrelevant individuals to protect their privacy efficiently. The method employs multi-object tracking to extract spatiotemporal motion features, enabling the automated recognition of irrelevant individuals. In addition, a trajectory association algorithm is employed to improve the precision of tracking from occlusion and blurring during motion. Additionally, as the field of irrelevant people privacy protection lacks data support, this paper constructs a dataset for further research, consisting of 101 685 irrelevant faces and 40 247 relevant faces. Through experimental validation on various datasets, the proposed method shows significant improvements in various metrics compared with the state-of-the-art approaches, with MOTA increasing by 3.2%, HOTA increasing by 13%, and the accuracy of individual irrelevant recognition reaching 96.78%.

Abstract The spread of disinformation on online social media has caused massive concern. Existing disinformation detection methods neglect the diverse compositional forms of tweets in real-life scenarios, making them less applicable and effective in social media settings. Meanwhile, these methods use pattern cues but overlook important aspects such as syntax, lexicon, and shallow visual semantics, and lack attention to factual content such as time, place, and person relay in both text and images, thus failing to fully explore features of disinformation and limiting detection accuracy. Furthermore, with the popularity of large language models (LLMs), the tweets generated by these models make the style of disinformation more subtle. Since existing datasets are mostly human-generated and lack style diversity, it results in weak detection capabilities of methods trained on these datasets. To address these challenges, a dual-feature adaptive framework for multimodal disinformation detection is proposed. The framework first using a similarity-based algorithm adaptively handles different tweet forms. It then enhances pattern features by bridging multimodal output from single-modal pretrained modal, and factual features are subsequently extracted using a zero-shot method based on a large vision language model. Finally, an expert network aggregates and reweights the dual-feature representation for tweets using an LLM-text detector in gating strategy. This paper also presents two multimodal disinformation datasets that include both LLM-generated and human-generated tweets reflecting real-world scenarios. The true tweets in datasets are diverse in style, while the fake tweets are more misleading. Experimentally verified, the proposed method outperforms baseline methods by an accuracy of 1.04% and 0.72% on typical datasets while also achieving a minimum accuracy drop of 0.65% and 0.87% on the proposed dataset.

The Motor Imagery (MI) decoding based on electroencephalogram (EEG), has promising applications. However, most current methods face two main issues: (1) They usually rely on convolutional neural networks to extract temporal features of MI signals without fully considering the brain’s functional connectivity during MI tasks. (2) They lack analysis and recognition of MI features slices and non-tasks slices within EEG signals, leading to poor generalization and robustness. To address these problems, we propose a novel deep learning model based on graph neural network to learn spatial features between multiple electrode channels and integrate the brain’s functional connectivity features. Additionally, it restructures time slices features segmented by the sliding time window algorithm to enhance MI temporal features in EEG signal. Therefor our model achieves the fusion of spatial and temporal features. To enhance the convergence effect of the model, we introduce electrode channel spatial positions as prior knowledge to initialize the parameters of the graph convolutional network parameters. Experimental evaluations on the publicly available EEG MI dataset from BCI Competition IV 2a show that our model achieves a four-class cross-session classification accuracy of 82.38%. Compared with other methods, our model yields the best results, demonstrating its superiority. Furthermore, the results indicate that the spatial feature obtained through our model bears resemblance to the brain functional connectivity patterns identified during MI tasks. To conclude, the fusion of spatial and temporal features with graph model shows the great application potential for EEG MI signals decoding and other EEG analysis.

The brain-computer interface (BCI) based on electroencephalogram (EEG) motor imagery (MI) decoding demonstrates promising application potential. However, the domain shift between training and testing data significantly impacts the model’s decoding efficacy. Domain adaption (DA) has been developed to address this problem recently. Nevertheless, existing DA methods have two limitations. One is that the extracted features are noisy, and the other is that they only align the distribution of features, which leads to limited generalization ability of the model. In this paper, we propose a novel smooth conditional domain adversarial training framework for solving the motor imagery decoding problem under domain shift. The framework uses interactive frequency convolution and channel attention mechanism as feature extractors to obtain effective features, and integrates smooth conditional domain adversarial training with batch spectral penalty to align the joint distribution of features and classes. At the same time, self-iterative training is implemented by generating pseudo-labels and selective outlier removal. Experimental results demonstrate that our proposed framework achieves 80.67% and 86.17% average accuracy in the BCI IV 2a and 2b respectively for cross-session experiments, achieving the best results compared with other methods, proving that the framework can improve the classification ability on the target domain while transferring effective features.

While marketing budget allocation has been studied for decades in traditional business, nowadays online business brings much more challenges due to the dynamic environment and complex decision-making process. In this paper, we present a novel unified framework for marketing budget allocation. By leveraging abundant data, the proposed data-driven approach can help us to overcome the challenges and make more informed decisions. In our approach, a semi-black-box model is built to forecast the dynamic market response and an efficient optimization method is proposed to solve the complex allocation task. First, the response in each market-segment is forecasted by exploring historical data through a semi-black-box model, where the capability of logit demand curve is enhanced by neural networks. The response model reveals relationship between sales and marketing cost. Based on the learned model, budget allocation is then formulated as an optimization problem, and we design efficient algorithms to solve it in both continuous and discrete settings. Several kinds of business constraints are supported in one unified optimization paradigm, including cost upper bound, profit lower bound, or ROI lower bound. The proposed framework is easy to implement and readily to handle large-scale problems. It has been successfully applied to many scenarios in Alibaba Group. The results of both offline experiments and online A/B testing demonstrate its effectiveness.

Many machine intelligence techniques are developed in E-commerce and one of the most essential components is the representation of IDs, including user ID, item ID, product ID, store ID, brand ID, category ID etc. The classical encoding based methods (like one-hot encoding) are inefficient in that it suffers sparsity problems due to its high dimension, and it cannot reflect the relationships among IDs, either homogeneous or heterogeneous ones. In this paper, we propose an embedding based framework to learn and transfer the representation of IDs. As the implicit feedbacks of users, a tremendous amount of item ID sequences can be easily collected from the interactive sessions. By jointly using these informative sequences and the structural connections among IDs, all types of IDs can be embedded into one low-dimensional semantic space. Subsequently, the learned representations are utilized and transferred in four scenarios: (i) measuring the similarity between items, (ii) transferring from seen items to unseen items, (iii) transferring across different domains, (iv) transferring across different tasks. We deploy and evaluate the proposed approach in Hema App and the results validate its effectiveness.

Existing works for extracting navigation objects from webpages focus on navigation menus, so as to reveal the information architecture of the site. However, web 2.0 sites such as social networks, e-commerce portals etc. are making the understanding of the content structure in a web site increasingly difficult. Dynamic and personalized elements such as top stories, recommended list in a webpage are vital to the understanding of the dynamic nature of web 2.0 sites. To better understand the content structure in web 2.0 sites, in this paper we propose a new extraction method for navigation objects in a webpage. Our method will extract not only the static navigation menus, but also the dynamic and personalized page-specific navigation lists. Since the navigation objects in a webpage naturally come in blocks, we first cluster hyperlinks into different blocks by exploiting spatial locations of hyperlinks, the hierarchical structure of the DOM-tree and the hyperlink density. Then we identify navigation objects from those blocks using the SVM classifier with novel features such as anchor text lengths etc. Experiments on real-world data sets with webpages from various domains and styles verified the effectiveness of our method.